In contrast to the single-sample-based models, the extended Gonfiantini & Zuppi model (Gonfiantini and Zuppi, 2003; Han et al., 2014) is a statistical approach.
In the extended Gonfiantini & Zuppi model, a representative system-wide value of the initial C data of DIC and can differ from that used in single-sample-based models.
Therefore, the extended Gonfiantini & Zuppi model usually avoids the effect of modern water components which might retain ‘bomb’ pulse signatures.
The geochemical mass-balance approach constructs an adjustment model that accounts for all the geochemical reactions known to occur along an aquifer flow path (Plummer et al., 1983; Wigley et al., 1978; Plummer et al., 1994; Plummer and Glynn, 2013), and includes, in addition to DIC, dissolved organic carbon (DOC) and methane (CH).
) — a quantity that is essential for estimation of radiocarbon age of DIC in groundwater.
The models/approaches most commonly used are grouped as follows: (1) single-sample-based models, (2) a statistical approach based on the observed (curved) relationship between C data for the aquifer, and (3) the geochemical mass-balance approach that constructs adjustment models accounting for all the geochemical reactions known to occur along a groundwater flow path.
This review discusses first the geochemical processes behind each of the single-sample-based models, followed by discussions of the statistical approach and the geochemical mass-balance approach.Finally, the applications, advantages and limitations of the three groups of models/approaches are discussed.The single-sample-based models constitute the prevailing use of C age, therefore the measurement data are easily available.These models have been shown to provide realistic radiocarbon ages in many studies.However, they usually are limited to simple carbonate aquifers and selection of model may have significant effects on of DIC in groundwater: Pearson's model, (Ingerson and Pearson, 1964; Pearson and White, 1967), Han & Plummer's model (Han and Plummer, 2013), the IAEA model (Gonfiantini, 1972; Salem et al., 1980), and Oeschger's model (Geyh, 2000).These four models include all processes considered in single-sample-based models, and can be used in different ranges of C values.